Modelling spatial heterogeneity in the haemodynamic response with implications for neuroimaging
Date of Award
Master of Science (Medicine)
Schools and Centres
Dr Edward Waters
Dr Ben Hui
Background: A foundational assumption of neuroimaging is the central volume principle (CVP): the mean transit time of oxygen particles in the brain equals the ratio of blood vessel volume to blood flow. Changes in mean transit time are expected to cause detectable changes in images produced by functional magnetic resonance imaging (fMRI). The CVP assumes a uniform distribution of transit times, but in fact blood vessel volumes are spatially heterogeneous. This thesis examines the implications of spatial heterogeneity for fMRI research.
Methods: An amended form of the CVP that accounts for spatial heterogeneity is developed and parameterised using empirical data. Implications of spatial heterogeneity and oxygen extraction for fMRI research are then examined using computer simulations.
Results: Spatial heterogeneity significantly reduces mean transit times; however, parameterisation of the model shows that, contrary to expectation, transit times might be uniformly distributed rather than heterogeneous. Nonetheless, computer simulations showed that common experimental designs are inadequate to detect clinically meaningful changes in transit time. Again, contrary to expectation, oxygen extraction is found to have no significant effect on mean transit time.
Conclusion: This thesis casts doubt on the degree to which spatial heterogeneity causes problems in neuroimaging, but nonetheless reaffirms that existing experimental designs are inadequate to detect significant changes in transit time. Further research on the assumption that significant changes in neuroimages are linked to changes in transit time is required.
Gore, R. (2018). Modelling spatial heterogeneity in the haemodynamic response with implications for neuroimaging (Master of Science (Medicine)). University of Notre Dame Australia. https://researchonline.nd.edu.au/theses/195